Tissue-Speci ﬁ c Macrophage Responses to Remote Injury Impact the Outcome

of Subsequent Local Immune Challenge

Friedrich Felix Hoyer, Kamila Naxerova, Maximilian J. Schloss, Maarten Hulsmans, Anil V.

Nair, Partha Dutta, David M. Calcagno, Fanny Herisson, Atsushi Anzai, Yuan Sun, Gregory Wojtkiewicz, David Rohde, Vanessa Frodermann, Katrien Vandoorne, Gabriel Courties, Yoshiko Iwamoto, Christopher S. Garris, David L. Williams, Sylvie Breton, Dennis Brown, Michael Whalen, Peter Libby, Mikael J. Pittet, Kevin R.

King, Ralph Weissleder, Filip K. Swirski, and Matthias Nahrendorf

70%

SSC-A MHCII - Alexa 700

SSC-A

FSC-A

CD64 - Alexa 647 SiglecF - BV421

CD11b - APC-Cy7

CD11c - BV510

SSC-A MHCII - Alexa 700

SSC-A

FSC-A

CD64 - Alexa 647 SiglecF - BV421

CD11b - APC-Cy7

CD11c - BV510

SSC-A MHCII - Alexa 700

SSC-A

FSC-A

CD64 - Alexa 647 SiglecF - BV421

CD11b - APC-Cy7

CD11c - BV510

SSC-A MHCII - Alexa 700

SSC-A

FSC-A

CD64 - Alexa 647 SiglecF - BV421

CD11b - APC-Cy7 gated on viable, single cells

CD11b - APC-Cy7

Lin + Ly6g - PE

FSC-A

CD64 - Alexa 647

Ly6c - FITC

CD64 - Alexa 647

Ly6c - FITC

CD64 - Alexa 647

Ly6c - FITC

F4/80 - PE-Cy7

Stroke gated on viable, single cells

Lin + Ly6g - PE CD64 - Alexa 647 F4/80 - PE-Cy7

CD11b - APC-Cy7

CD45 - BV711

Ly6c - FITC FSC-A CD11b - APC-Cy7

67% 95% 98% 98%

Control

Lin + Ly6g - PE CD64 - Alexa 647 F4/80 - PE-Cy7

CD11b - APC-Cy7

CD45 - BV711

Ly6c - FITC FSC-A CD11b - APC-Cy7

62% 93% 99% 98%

Lin + Ly6g - PE CD64 - Alexa 647 F4/80 - PE-Cy7

CD11b - APC-Cy7

CD45 - BV711

Ly6c - FITC FSC-A CD11b - APC-Cy7

76% 84% 98% 99%

gated on viable, single cells

CD64 - Alexa 647

12%

CD64 - Alexa 647

15%

CD64 - Alexa 647

CD11b - APC-Cy7

F4/80 - PE-Cy7

CD45 - PerCP-Cy5

Lin + Ly6g - PE

SSC-A

CD64 - Alexa 647

gated on viable, single cells

CLP

CD45 - PerCP-Cy5

Lin + Ly6g - PE

SSC-A

CD64 - Alexa 647

CD11b - APC-Cy7

CD64 - Alexa 647

CD11b - APC-Cy7

CD64 - Alexa 647

CD11b - APC-Cy7

CD64 - Alexa 647

CD11b - APC-Cy7

F4/80 - PE-Cy7

14%

86% 89%

Figure S1

gated on viable, single cells

A B

Figure S1. Flow cytometry gating strategy for all organs and conditions, Related to Figure 1-4.

Gating for macrophages in (A) lung, (B) heart, (C) brain, (D) kidney and (E) liver followed the

Immgen guidelines whenever available. By design, the gates include both, resident and newly

recruited macrophages. 

MI Con 15

Recruited renal macrophages (%) Monocytes (102)/ mg liver

Neutrophils (102) mg liver

** Recruited renal macrophages (%)

* Interstitial macrophages (102)/ mg lung

gated on CD11c^low MHCII^high B

Figure S2 Cx3cr1-YFP^CreER/+

R26-tdTomato^tdTomato/+

100 50

tdTomatohigh Ly6c monocytes (%)

day 1 21

Digested lung BAL BAL d4 after MI Interstitial macrophages % of CD64+ cells

F C

Cx3cr1CreER/+ R26tdTomato/+wildtype 6% 63% 93% 0%25%

15% 45% 12% 93% 0.1%

12% 44% 15% 87% 0.03%

Cx3cr1CreER/+ R26tdTomato/+ CD11c

SSC-A MHCII

YFPhigh alveolar macrophages (%)

Con

gated on alveolar macrophages

0.06%

BrdU+ alveolar macrophages (%)

BrdU+ alveolar macrophages (%) Con MI

flow BrdU pulse for 14d via drinking waterMI d1d14d15d19

Monocytes (103)/ mg lung 1.5

3 ****

Neutrophils (103)/ mg lung 1.5

Monocytes (103)/ mg lung 0.7 1.4

Neutrophils (103)/ mg lung 1.5

Neutrophils (103) mg lung

days

Monocytes (101)/ mg heart Stroke Neutrophils (102)/ mg heart

Con 1 4 Monocytes (102)/ mg heart

CLP

***

Con 1 4 71428 63 8

4 Neutrophils (102)/ mg heart

CLP

Monocytes (102)/ mg liver 5 2.5 Neutrophils (102) mg liver

************

Con 1 4 7 1421 Monocytes (102)/ mg liver 1

0.5

***

5 2.5 Neutrophils (102) mg liver **

Con 1 4 7 Stroke

Stroke

Con 1471428 day

CD11bhigh macro phages (103)/ mg liver

1.5

YFPhigh recruited liver macrophages (%)

Recruited CD45.1 macrophages (%) 10

YFPhigh tdTomatolow macrophages (%)

***

Gated on liver macrophages ****

Renal macrophages in SG2M % Con

Monocytes (101)/ mg kidney

CLP

**** *

15 7.5

Neutrophils (101)/ mg kidney

Con 134 714 2863

**** ****

CLP

7 ********

3.5 Monocytes (101)/ mg kidney

******

4 Neutrophils (101)/ mg kidney

Stroke 4 2 Monocytes (101)/ mg kidney

Stroke Neutrophils (101)/ mg kidney

Con 1 4 7 day

YFPYFPhigh tdTomatolow renal macrophages (%)30

Renal macrophages in SG2M %

Con 134 7 1428 63

CD31 F4/80

Control CLP d4

Neutrophils (101)/ mg brain

MI 1.2

2.4

Monocytes (101)/ mg brain ****

Neutrophils (101)/ mg brain 2 CD45high macrophages (101)/ mg brain

W ₆

CD45high macrophages (101)/ mg brain

Figure S2. Myeloid cells and macrophage fate mapping after remote injury in the lung, heart, liver and kidney, Related to Figure 1.

(A) Interstitial macrophages after injury. Control n=19-28, MI n=4-12, stroke n=8-13, CLP n=5-10 biological replicates. (B) Fate mapping in Cx3cr1

^CreER/+

R26

^tdTomato/+

mice and quantification of fluorescent proteins in blood monocytes, day 1 n=9, day 21 n=9. (C) Fate mapping for alveolar macrophages in Cx3cr1

^CreER/+

R26

^tdTomato/+

mice after MI and CLP and naive Cx3cr1

^CreER/

R26

^tdTomato/+

controls. Control n=8, MI n=10, CLP n=6 biological replicates. (D) Fate mapping experiments for alveolar macrophages in CD45.1/2 parabionts. MI was induced in DC45.2 mouse, whose lung was then investigated. MI n=2 pairs, naive control n=2 pairs. (E)

Proliferation of alveolar macrophages by BrdU flow cytometry. Left: BrdU was injected i.p. three hours before sacrifice, n=8-10. Right: BrdU pulse chase experiment, n=6-8. (F) FACS of

bronchoalveolar lavage (BAL) in naive controls and day 4 after MI. Naive controls (digested) n=4, controls BAL n=5, MI n=5 biological replicates. (G) Lung monocytes and neutrophils after MI, n=4-30; after stroke, n=8-19; and after CLP, n=5-21 biological replicates. (H) Fate mapping for cardiac macrophages in Cx3cr1

^CreER/+

R26

^tdTomato/+

mice after CLP and naive Cx3cr1

^CreER/

R26

^tdTomato/+

controls. Control n=8, CLP n=12 biological replicates. (I) Fate mapping for cardiac macrophages in CD45.1/2 parabionts after CLP. MI n=6 pairs, naive control n=6 pairs. (J) Cardiac monocytes and neutrophils after stroke, n=4-18 biological replicates per time point; and after CLP, n=5-20. (K) Cell cycle analysis of cardiac macrophages after CLP, naive control n=5, CLP n=4 biological replicates. (L) Fluorescence images of hepatic macrophages at baseline and 4 days after CLP. Scale bar, 50 µm. (M) Hepatic monocytes and neutrophils after MI, n=4-31;

after stroke, n=5-19; and after CLP, n=5-16. (N) Hepatic CD11b

^high

macrophages, control n=16-32, MI n=4-12, stroke n=5-13, CLP n=5-6 biological replicates. (O) Fate mapping for hepatic macrophages in Cx3cr1

^CreER/+

R26

^tdTomato/+

mice after CLP or MI and naive Cx3cr1

^CreER/

R26

^tdTomato/+

controls. Control n=7, MI n=9, CLP n=9 biological replicates. (P) Histology of renal macrophages after CLP, Scale bar, 50 µm. (Q) Proliferation of renal macrophages by Ki-67 and Propidium iodide (PI) flow cytometry on d7 after MI, n=7-12. (R) Recruitment of renal

macrophages after MI assessed by parabiosis, n=5-8. (S) Cell cycle analysis in renal

macrophage on d7 after CLP, n=4-6 biological replicates. (T) Recruitment of renal macrophages after CLP assessed by parabiosis, n=10-11; and (U) after MI and CLP in Cx3cr1

^CreER/

R26

^tdTomato/+

mice, n=3-6. (V) Renal monocytes and neutrophils after MI, n=4-34; after stroke, n=4-16; and after CLP, n=5-29. (W) Brain CD45

^high

macrophages after MI and CLP, n=8-14 biological replicates per time point. (X) Monocytes and neutrophils in the brain after MI, n=8 and after CLP, n=8-14. (Y) Fate mapping for microglia in Cx3cr1

^CreER/+

R26

^tdTomato/+

mice after CLP, after MI and in and naive Cx3cr1

^CreER/+

R26

^tdTomato/+

controls. Control n=5, MI n=8, CLP n=6 biological replicates. Data are mean ± s.e.m. P<0.05, P<0.01, P<0.001, ****P<0.0001.

One-way ANOVA with Dunnett’s multiple comparison test was applied for normally distributed

data. 

Figure S3

Enrichment plot: Goldrath antigen response

Brain Lung

Enrichment plot: Goldrath antigen response

Brain Heart

Enrichment plot: Goldrath antigen response

Brain Kidney

Enrichment plot: Goldrath antigen response

Brain Liver

Enrichment plot: Galindo immune response to enterotoxin

Brain Lung

Enrichment plot: Galindo immune response to enterotoxin

Brain Heart

Enrichment plot: Galindo immune response to enterotoxin

Brain Kidney

Enrichment plot: Galindo immune response to enterotoxin

Brain Liver

Brain.Control Brain.MI Brain.Sepsis Heart.Control Heart.Sepsis Heart.Stroke Kidney.Control Kidney.MI Kidney.Sepsis Kidney.Stroke Liver.Control Liver.MI Liver.Sepsis Liver.Stroke Lung.Control Lung.MI Lung.Sepsis Lung.Stroke

Condition

Brain control Brain MI Brain CLP Heart CLP Heart stroke Kidney control Kidney MI Kidney CLP Kidney stroke Livercontrol Liver MI Liver CLP Liver stroke Lung control Lung MI Lung CLP Lung stroke

Brain control Brain MI Brain CLP Heart control Heart CLP Heart stroke Kidney control Kidney MI Kidney CLP Kidney stroke Liver control Liver MI Liver CLP Liver stroke Lung control Lung MI Lung CLP Lung stroke

NES -2.45

Brain control Brain MI Brain CLP Heart CLP Heart stroke Kidney control Kidney MI Kidney CLP Kidney stroke Livercontrol Liver MI Liver CLP Liver stroke Lung control Lung MI Lung CLP Lung stroke

Heart control

Figure S3. Organ specific macrophage expression changes after injury, Related to Figure 2 and 3.

(A) Gene set enrichment analysis of the gene set “GOLDRATH_ANTIGEN_RESPONSE“ in brain microglia vs. all other organs. FDRs are calculated by gene set permutation. (B) Gene set enrichment analysis of the gene set “GALINDO_IMMUNE_RESPONSE_TO_ENTEROTOXIN“

in brain microglia vs. all other organs. (C) Expression (RMA log2 signal) of Cxcl2, (D) Ccl12, (E)

Tnfsf9 across the entire data set. 

Interstitial macrophages (102)/ mg lung

2.5 Interstitial macrophages (102)/ mg lung

2.5 5

gated on CD11chigh, MHCIIhigh

Alveolar

Egfr expression fold-change over control

LysM Egfr^+/+Sham n=10 LysM Egfr^-/-Sham n=12

Control CloLip MI MI + CloLip

LysM Egfr^-/-Sham thoracotomy + Pneumonia LysM Egfr^+/+ Sham thoracotomy + Pneumonia Naive LysM Egfr^+/+

Figure S4. Alveolar macrophages after MI, Related to Figure 5.

(A) Bioluminescence signal 2 days after instillation of bioluminescent bacteria in naive controls, on day 6 after MI and day 6 after sham thoracotomy surgery. Control n=39, MI n=23, sham thoracotomy n=18. (B) Flow plots of lung interstitial macrophages in wildtype and Ccr2

^-/-

mice after MI. (C) Quantification of lung interstitial macrophages in wildtype and Ccr2

^-/-

mice after MI, n=4-5. (D) Ccr2

^-/-

mice were subjected to MI and received bioluminescent streptococcus 4 days later, n=8-15. (E) Alveolar macrophages were depleted via intra-tracheal instillation of

clodronate liposome. Flow plots shows macrophages in naive control and after liposome treatment. (F) Quantification of alveolar macrophages over time after clodronate depletion, n=2-4 biological replicates per time point. (G) Flow cytometry for interstitial lung macrophages after intratracheal clodronate liposome instillation. Con n=4, Clolip n=4 biological replicates. (H) Bioluminescence signal 2 days after instillation of bioluminescent bacteria in naive controls, controls that received intratracheal clodronate liposomes, and on day 6 after MI with and without intratracheal clodronate. Naive control n=27, control with clodronate n=12, MI n=11, MI with clodronate n=12. (I) Expression of Egfr in flow sorted alveolar macrophages, interstitial lung macrophages and plasmacytoid dendritic cells. Naive control n=5-6 biological replicates, MI n=6, Pneumonia n=6. (J) Survival analysis in mice with sham thoracotomy. Sham LysM Egfr

^-/-

n=12, LysM Egfr

^+/+

n=12. (K) Gene expression in lung tissue harvested from mice that

underwent sham thoracotomy control procedures and induction of pneumonia. Naive LysM Egfr

^+/+

n=x, Sham LysM Egfr

^-/-

n=8, LysM Egfr

^+/+

n=7. Data are mean ± s.e.m. P<0.05,*

P<0.01, *P<0.001, ****P<0.0001. Unpaired t-test or ordinary one-way ANOVA with Dunnett’s

multiple comparison test was used for analysis. 

Figure S5. Gene expression changes in lung and heart macrophages, Related to Figure 5.

(A) Experimental outline. (B) Gene expression in alveolar macrophages, (C) interstitial lung macrophages, (D) plasmacytoid dendritic cells, (E) neutrophils in naive LysM Egfr

^+/+

n=2-4, LysM Egfr

^+/+

with MI and pneumonia n=4-8, LysM Egfr

^-/-

with MI and pneumonia n=4-8 biological replicates. (F) Gene expression in alveolar macrophages after sham thoracotomy control

procedure and pneumonia. LysM Egfr

^+/+

naive n=4, LysM Egfr

^+/+

Sham thoracotomy +

Pneumonia n=4-8, LysM Egfr

^-/-

Sham thoracotomy + Pneumonia n=4-8 biological replicates.

(G) Bioluminescence imaging after 10µg of intratracheal IFNɣ and intratracheal streptococcus pneumoniae instillation. Data are mean ± s.e.m. *P<0.05. Unpaired t-test or ordinary one-way ANOVA with Dunnett’s multiple comparison test was used for analysis. 

LysM Egfr^-/-Sham thoracotomy + Pneumonia LysM Egfr^+/+ Sham thoracotomy + Pneumonia LysM Egfr^+/+naive

4 0 days

Sham Lung infection FACS Pneumonia

Cd11c^Cre/+Ifngr1^flox/flox

KLRB1C NK cells

YFP-IFNg

modal

KLRB1C NK cells KLRB1C NK cells

Flow

NK cells (102)/ mg cardiac tissue 3 6 ****

Con MI

Con MI NK cells (103)/ ml

CD8 T cells CD24 CD103 CD8 T cells CD24 CD103 DCs

DCs

CD8 T cells CD4 T cells CD11b Myeloid cells

YFP-IFNg

modal

C

Figure S6. Disrupted IFNɣ signaling abrogates lung protection after MI, Related to Figure 5.

(A,B) IFNɣ expression of NK cells in heart, blood, and lung in control and YFP-IFNɣ reporter mice. (C) No change in IFNɣ expression in indicated cells in heart, blood and lung. (D) Gating and (E) quantification of cardiac NK cells in control mice and on day 4 after MI. Control n=7, MI n=9. (F) Gating and (G) quantification of blood NK cells of control mice on day 4 after MI.

Control n=7, MI n=9 biological replicates. (H) Gating and (I) quantification of lung NK cells in

control mice and on day 4 after MI. Control n=7, MI n=9 biological replicates. (J)

Dans le document Tissue-specific macrophage responses to remote injury impact the outcome of subsequent local immune challenge (Page 26-37)

Tissue-Speci ﬁ c Macrophage Responses to Remote Injury Impact the Outcome

of Subsequent Local Immune Challenge

Friedrich Felix Hoyer, Kamila Naxerova, Maximilian J. Schloss, Maarten Hulsmans, Anil V.

King, Ralph Weissleder, Filip K. Swirski, and Matthias Nahrendorf

Figure S1. Flow cytometry gating strategy for all organs and conditions, Related to Figure 1-4.

Gating for macrophages in (A) lung, (B) heart, (C) brain, (D) kidney and (E) liver followed the

Immgen guidelines whenever available. By design, the gates include both, resident and newly

recruited macrophages.

Figure S2. Myeloid cells and macrophage fate mapping after remote injury in the lung, heart, liver and kidney, Related to Figure 1.

(A) Interstitial macrophages after injury. Control n=19-28, MI n=4-12, stroke n=8-13, CLP n=5-10 biological replicates. (B) Fate mapping in Cx3cr1

R26

mice and quantification of fluorescent proteins in blood monocytes, day 1 n=9, day 21 n=9. (C) Fate mapping for alveolar macrophages in Cx3cr1

R26

mice after MI and CLP and naive Cx3cr1

R26

controls. Control n=8, MI n=10, CLP n=6 biological replicates. (D) Fate mapping experiments for alveolar macrophages in CD45.1/2 parabionts. MI was induced in DC45.2 mouse, whose lung was then investigated. MI n=2 pairs, naive control n=2 pairs. (E)

Proliferation of alveolar macrophages by BrdU flow cytometry. Left: BrdU was injected i.p. three hours before sacrifice, n=8-10. Right: BrdU pulse chase experiment, n=6-8. (F) FACS of

R26

mice after CLP and naive Cx3cr1

R26

after stroke, n=5-19; and after CLP, n=5-16. (N) Hepatic CD11b

macrophages, control n=16-32, MI n=4-12, stroke n=5-13, CLP n=5-6 biological replicates. (O) Fate mapping for hepatic macrophages in Cx3cr1

R26

mice after CLP or MI and naive Cx3cr1

R26

controls. Control n=7, MI n=9, CLP n=9 biological replicates. (P) Histology of renal macrophages after CLP, Scale bar, 50 µm. (Q) Proliferation of renal macrophages by Ki-67 and Propidium iodide (PI) flow cytometry on d7 after MI, n=7-12. (R) Recruitment of renal

macrophages after MI assessed by parabiosis, n=5-8. (S) Cell cycle analysis in renal

macrophage on d7 after CLP, n=4-6 biological replicates. (T) Recruitment of renal macrophages after CLP assessed by parabiosis, n=10-11; and (U) after MI and CLP in Cx3cr1

R26

mice, n=3-6. (V) Renal monocytes and neutrophils after MI, n=4-34; after stroke, n=4-16; and after CLP, n=5-29. (W) Brain CD45

macrophages after MI and CLP, n=8-14 biological replicates per time point. (X) Monocytes and neutrophils in the brain after MI, n=8 and after CLP, n=8-14. (Y) Fate mapping for microglia in Cx3cr1

R26

mice after CLP, after MI and in and naive Cx3cr1

R26

controls. Control n=5, MI n=8, CLP n=6 biological replicates. Data are mean ± s.e.m. *P<0.05, **P<0.01, ***P<0.001, ****P<0.0001.

One-way ANOVA with Dunnett’s multiple comparison test was applied for normally distributed

data.

Figure S3

Figure S3. Organ specific macrophage expression changes after injury, Related to Figure 2 and 3.

(A) Gene set enrichment analysis of the gene set “GOLDRATH_ANTIGEN_RESPONSE“ in brain microglia vs. all other organs. FDRs are calculated by gene set permutation. (B) Gene set enrichment analysis of the gene set “GALINDO_IMMUNE_RESPONSE_TO_ENTEROTOXIN“

in brain microglia vs. all other organs. (C) Expression (RMA log2 signal) of Cxcl2, (D) Ccl12, (E)

Tnfsf9 across the entire data set.

Figure S4. Alveolar macrophages after MI, Related to Figure 5.

(A) Bioluminescence signal 2 days after instillation of bioluminescent bacteria in naive controls, on day 6 after MI and day 6 after sham thoracotomy surgery. Control n=39, MI n=23, sham thoracotomy n=18. (B) Flow plots of lung interstitial macrophages in wildtype and Ccr2

mice after MI. (C) Quantification of lung interstitial macrophages in wildtype and Ccr2

mice after MI, n=4-5. (D) Ccr2

mice were subjected to MI and received bioluminescent streptococcus 4 days later, n=8-15. (E) Alveolar macrophages were depleted via intra-tracheal instillation of

n=12, LysM Egfr

n=12. (K) Gene expression in lung tissue harvested from mice that

underwent sham thoracotomy control procedures and induction of pneumonia. Naive LysM Egfr

n=x, Sham LysM Egfr

n=8, LysM Egfr

n=7. Data are mean ± s.e.m. *P<0.05,

**P<0.01, ***P<0.001, ****P<0.0001. Unpaired t-test or ordinary one-way ANOVA with Dunnett’s

multiple comparison test was used for analysis.

Figure S5. Gene expression changes in lung and heart macrophages, Related to Figure 5.

(A) Experimental outline. (B) Gene expression in alveolar macrophages, (C) interstitial lung macrophages, (D) plasmacytoid dendritic cells, (E) neutrophils in naive LysM Egfr

n=2-4, LysM Egfr

with MI and pneumonia n=4-8, LysM Egfr

with MI and pneumonia n=4-8 biological replicates. (F) Gene expression in alveolar macrophages after sham thoracotomy control

procedure and pneumonia. LysM Egfr

naive n=4, LysM Egfr

Sham thoracotomy +

Pneumonia n=4-8, LysM Egfr

Sham thoracotomy + Pneumonia n=4-8 biological replicates.

(G) Bioluminescence imaging after 10µg of intratracheal IFNɣ and intratracheal streptococcus pneumoniae instillation. Data are mean ± s.e.m. *P<0.05. Unpaired t-test or ordinary one-way ANOVA with Dunnett’s multiple comparison test was used for analysis.

C

Figure S6. Disrupted IFNɣ signaling abrogates lung protection after MI, Related to Figure 5.

Control n=7, MI n=9 biological replicates. (H) Gating and (I) quantification of lung NK cells in

control mice and on day 4 after MI. Control n=7, MI n=9 biological replicates. (J)

recruited macrophages. 

controls. Control n=5, MI n=8, CLP n=6 biological replicates. Data are mean ± s.e.m. P<0.05, P<0.01, P<0.001, ****P<0.0001.

data. 

Tnfsf9 across the entire data set. 

n=7. Data are mean ± s.e.m. P<0.05,*

P<0.01, *P<0.001, ****P<0.0001. Unpaired t-test or ordinary one-way ANOVA with Dunnett’s

multiple comparison test was used for analysis. 

(G) Bioluminescence imaging after 10µg of intratracheal IFNɣ and intratracheal streptococcus pneumoniae instillation. Data are mean ± s.e.m. *P<0.05. Unpaired t-test or ordinary one-way ANOVA with Dunnett’s multiple comparison test was used for analysis. 